[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Hard to debug with (load nomessage)
From: |
Järneström Jonas |
Subject: |
Hard to debug with (load nomessage) |
Date: |
Wed, 20 Mar 2002 12:46:32 +0100 (MET) |
During loading of my main library I got
"File mode specification error: end-of-file"
I checked *messages* and last item was a completed load of my
main-library, i.e. "file....done"
Not seeing any other file to suspect, I still (wrongly) suspected
my main library.
My main library does a lot of silent loading, i.e. with the nomessage
switch on. Nomessage in this context means "no mess in echo area and
consequently no mess in *messages*. I used nomessage because I didn't
want to show the numerous loading messages to the end-user, since they
are many and have no relevance to him/her. For debug purposes, I
surely would like to keep trace printouts of any loading that
takes place.
Anyway, the error was corrupted lisp code in one of my silent load
files, and I didnt realize that using nomessage suppresses both echo
area and *messages* (i.e. that they are virtually the same thing).
Since I had 20 load files to chose from, I couldnt come up with a method
to make the offending file visible without removing the nomessage
option when calling (load).
My point is thus that using (load nomessage) if you do a lot of
loading in your application is potentially dangerous from a debugging
point of view. I think this should be stated in the on-line doc for (load)
Further, I would like to see a mechanism that permits silent loading
with regard to echo area/end-user, but still leaves a trace message in
*messages* (or some other trace buffer). I managed to get a similar
effect by calling (load) with message and immediately after call
(message nil) to clear the echo area. But nicer would be if that
behaviour was controllable from within (load).
- Hard to debug with (load nomessage),
Järneström Jonas <=